Figure out why different farkas certificates generate same cut their respective disjunctive terms

[spkelle2]

[spkelle2]

Ok. As to why the RHS is changing between two instances but the first's instances farkas coefficients and disjunctions are applied to the second and we still end up with the same cut being generated for all disjunctive terms: I'm guessing the RHS is changing on rows that remain slack, thus the basis cone is unchanged, resulting in the same cuts being generated despite the problems being different in their constraint bounds

[spkelle2]

The solution should probably be to analyze the problem. Look at if it changed much. If so, generate cuts. If not, reuse them.